Ajeya Cotra on how Artificial Intelligence Could Cause Catastrophe
Future of Life Institute Podcast
00:00
How to Reduce the Failure Rate of Language Models
You could imagine that language models or more developed the future models will have to have a very low failure rate in order for their plans to be deployed. You wouldn't be worried that these silly mistakes when they pop up, they could ruin their whole plan of a model. But unlike self-driving cars a thing you can do with language models is just run them over and over again. So I expect you will both increase the kind of raw first cut accuracy per step and do a number of tricks to make things more robust.
Transcript
Play full episode
